hikopensource/DAVAR-Lab-OCR

Text-layout inference

Closed this issue · 3 comments

Thank you very much for your great work to merge CNN and NLP.
Can you please share your test_pipeline during inference or prediction? As the current pipeline is for evaluation.

The test pipeline code will be published soon.

Well, i think what you mean by "test_pipeline" is re-writing codes to preprocess the image (resize, pad, normalize...) and text (tokenize...) separately?
However, all these code are already in the code. And it's more convenient to use the current pipeline.

If you still have any doubt, feel free to post.

The test pipeline code will be published soon.

The code has been updated in e14ed62